| 1 | //===- TokensTest.cpp -----------------------------------------------------===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | |
| 9 | #include "clang/Tooling/Syntax/Tokens.h" |
| 10 | #include "clang/AST/ASTConsumer.h" |
| 11 | #include "clang/AST/Expr.h" |
| 12 | #include "clang/Basic/Diagnostic.h" |
| 13 | #include "clang/Basic/DiagnosticIDs.h" |
| 14 | #include "clang/Basic/DiagnosticOptions.h" |
| 15 | #include "clang/Basic/FileManager.h" |
| 16 | #include "clang/Basic/FileSystemOptions.h" |
| 17 | #include "clang/Basic/LLVM.h" |
| 18 | #include "clang/Basic/LangOptions.h" |
| 19 | #include "clang/Basic/SourceLocation.h" |
| 20 | #include "clang/Basic/SourceManager.h" |
| 21 | #include "clang/Basic/TokenKinds.def" |
| 22 | #include "clang/Basic/TokenKinds.h" |
| 23 | #include "clang/Frontend/CompilerInstance.h" |
| 24 | #include "clang/Frontend/FrontendAction.h" |
| 25 | #include "clang/Frontend/Utils.h" |
| 26 | #include "clang/Lex/Lexer.h" |
| 27 | #include "clang/Lex/PreprocessorOptions.h" |
| 28 | #include "clang/Lex/Token.h" |
| 29 | #include "clang/Tooling/Tooling.h" |
| 30 | #include "llvm/ADT/ArrayRef.h" |
| 31 | #include "llvm/ADT/IntrusiveRefCntPtr.h" |
| 32 | #include "llvm/ADT/STLExtras.h" |
| 33 | #include "llvm/ADT/StringRef.h" |
| 34 | #include "llvm/Support/FormatVariadic.h" |
| 35 | #include "llvm/Support/MemoryBuffer.h" |
| 36 | #include "llvm/Support/VirtualFileSystem.h" |
| 37 | #include "llvm/Support/raw_os_ostream.h" |
| 38 | #include "llvm/Support/raw_ostream.h" |
| 39 | #include "llvm/Testing/Annotations/Annotations.h" |
| 40 | #include "llvm/Testing/Support/SupportHelpers.h" |
| 41 | #include <cassert> |
| 42 | #include <cstdlib> |
| 43 | #include <gmock/gmock.h> |
| 44 | #include <gtest/gtest.h> |
| 45 | #include <memory> |
| 46 | #include <optional> |
| 47 | #include <ostream> |
| 48 | #include <string> |
| 49 | |
| 50 | using namespace clang; |
| 51 | using namespace clang::syntax; |
| 52 | |
| 53 | using llvm::ValueIs; |
| 54 | using ::testing::_; |
| 55 | using ::testing::AllOf; |
| 56 | using ::testing::Contains; |
| 57 | using ::testing::ElementsAre; |
| 58 | using ::testing::Field; |
| 59 | using ::testing::IsEmpty; |
| 60 | using ::testing::Matcher; |
| 61 | using ::testing::Not; |
| 62 | using ::testing::Pointee; |
| 63 | using ::testing::StartsWith; |
| 64 | |
| 65 | namespace { |
| 66 | // Checks the passed ArrayRef<T> has the same begin() and end() iterators as the |
| 67 | // argument. |
| 68 | MATCHER_P(SameRange, A, "" ) { |
| 69 | return A.begin() == arg.begin() && A.end() == arg.end(); |
| 70 | } |
| 71 | |
| 72 | Matcher<TokenBuffer::Expansion> |
| 73 | IsExpansion(Matcher<llvm::ArrayRef<syntax::Token>> Spelled, |
| 74 | Matcher<llvm::ArrayRef<syntax::Token>> Expanded) { |
| 75 | return AllOf(matchers: Field(field: &TokenBuffer::Expansion::Spelled, matcher: Spelled), |
| 76 | matchers: Field(field: &TokenBuffer::Expansion::Expanded, matcher: Expanded)); |
| 77 | } |
| 78 | // Matchers for syntax::Token. |
| 79 | MATCHER_P(Kind, K, "" ) { return arg.kind() == K; } |
| 80 | MATCHER_P2(HasText, Text, SourceMgr, "" ) { |
| 81 | return arg.text(*SourceMgr) == Text; |
| 82 | } |
| 83 | /// Checks the start and end location of a token are equal to SourceRng. |
| 84 | MATCHER_P(RangeIs, SourceRng, "" ) { |
| 85 | return arg.location() == SourceRng.first && |
| 86 | arg.endLocation() == SourceRng.second; |
| 87 | } |
| 88 | |
| 89 | class TokenCollectorTest : public ::testing::Test { |
| 90 | public: |
| 91 | /// Run the clang frontend, collect the preprocessed tokens from the frontend |
| 92 | /// invocation and store them in this->Buffer. |
| 93 | /// This also clears SourceManager before running the compiler. |
| 94 | void recordTokens(llvm::StringRef Code) { |
| 95 | class RecordTokens : public ASTFrontendAction { |
| 96 | public: |
| 97 | explicit RecordTokens(TokenBuffer &Result) : Result(Result) {} |
| 98 | |
| 99 | bool BeginSourceFileAction(CompilerInstance &CI) override { |
| 100 | assert(!Collector && "expected only a single call to BeginSourceFile" ); |
| 101 | Collector.emplace(args&: CI.getPreprocessor()); |
| 102 | return true; |
| 103 | } |
| 104 | void EndSourceFileAction() override { |
| 105 | assert(Collector && "BeginSourceFileAction was never called" ); |
| 106 | Result = std::move(*Collector).consume(); |
| 107 | Result.indexExpandedTokens(); |
| 108 | } |
| 109 | |
| 110 | std::unique_ptr<ASTConsumer> |
| 111 | CreateASTConsumer(CompilerInstance &CI, StringRef InFile) override { |
| 112 | return std::make_unique<ASTConsumer>(); |
| 113 | } |
| 114 | |
| 115 | private: |
| 116 | TokenBuffer &Result; |
| 117 | std::optional<TokenCollector> Collector; |
| 118 | }; |
| 119 | |
| 120 | constexpr const char *FileName = "./input.cpp" ; |
| 121 | FS->addFile(Path: FileName, ModificationTime: time_t(), Buffer: llvm::MemoryBuffer::getMemBufferCopy(InputData: "" )); |
| 122 | // Prepare to run a compiler. |
| 123 | if (!Diags->getClient()) |
| 124 | Diags->setClient(client: new IgnoringDiagConsumer); |
| 125 | std::vector<const char *> Args = {"tok-test" , "-std=c++03" , "-fsyntax-only" , |
| 126 | FileName}; |
| 127 | CreateInvocationOptions CIOpts; |
| 128 | CIOpts.Diags = Diags; |
| 129 | CIOpts.VFS = FS; |
| 130 | auto CI = createInvocation(Args, Opts: std::move(CIOpts)); |
| 131 | assert(CI); |
| 132 | CI->getFrontendOpts().DisableFree = false; |
| 133 | CI->getPreprocessorOpts().addRemappedFile( |
| 134 | From: FileName, To: llvm::MemoryBuffer::getMemBufferCopy(InputData: Code).release()); |
| 135 | CompilerInstance Compiler(std::move(CI)); |
| 136 | Compiler.setDiagnostics(Diags.get()); |
| 137 | Compiler.setFileManager(FileMgr.get()); |
| 138 | Compiler.setSourceManager(SourceMgr.get()); |
| 139 | |
| 140 | this->Buffer = TokenBuffer(*SourceMgr); |
| 141 | RecordTokens Recorder(this->Buffer); |
| 142 | ASSERT_TRUE(Compiler.ExecuteAction(Recorder)) |
| 143 | << "failed to run the frontend" ; |
| 144 | } |
| 145 | |
| 146 | /// Record the tokens and return a test dump of the resulting buffer. |
| 147 | std::string collectAndDump(llvm::StringRef Code) { |
| 148 | recordTokens(Code); |
| 149 | return Buffer.dumpForTests(); |
| 150 | } |
| 151 | |
| 152 | // Adds a file to the test VFS. |
| 153 | void addFile(llvm::StringRef Path, llvm::StringRef Contents) { |
| 154 | if (!FS->addFile(Path, ModificationTime: time_t(), |
| 155 | Buffer: llvm::MemoryBuffer::getMemBufferCopy(InputData: Contents))) { |
| 156 | ADD_FAILURE() << "could not add a file to VFS: " << Path; |
| 157 | } |
| 158 | } |
| 159 | |
| 160 | /// Add a new file, run syntax::tokenize() on the range if any, run it on the |
| 161 | /// whole file otherwise and return the results. |
| 162 | std::vector<syntax::Token> tokenize(llvm::StringRef Text) { |
| 163 | llvm::Annotations Annot(Text); |
| 164 | auto FID = SourceMgr->createFileID( |
| 165 | Buffer: llvm::MemoryBuffer::getMemBufferCopy(InputData: Annot.code())); |
| 166 | // FIXME: pass proper LangOptions. |
| 167 | if (Annot.ranges().empty()) |
| 168 | return syntax::tokenize(FID, SM: *SourceMgr, LO: LangOptions()); |
| 169 | return syntax::tokenize( |
| 170 | FR: syntax::FileRange(FID, Annot.range().Begin, Annot.range().End), |
| 171 | SM: *SourceMgr, LO: LangOptions()); |
| 172 | } |
| 173 | |
| 174 | // Specialized versions of matchers that hide the SourceManager from clients. |
| 175 | Matcher<syntax::Token> HasText(std::string Text) const { |
| 176 | return ::HasText(gmock_p0: Text, gmock_p1: SourceMgr.get()); |
| 177 | } |
| 178 | Matcher<syntax::Token> RangeIs(llvm::Annotations::Range R) const { |
| 179 | std::pair<SourceLocation, SourceLocation> Ls; |
| 180 | Ls.first = SourceMgr->getLocForStartOfFile(FID: SourceMgr->getMainFileID()) |
| 181 | .getLocWithOffset(Offset: R.Begin); |
| 182 | Ls.second = SourceMgr->getLocForStartOfFile(FID: SourceMgr->getMainFileID()) |
| 183 | .getLocWithOffset(Offset: R.End); |
| 184 | return ::RangeIs(gmock_p0: Ls); |
| 185 | } |
| 186 | |
| 187 | /// Finds a subrange in O(n * m). |
| 188 | template <class T, class U, class Eq> |
| 189 | llvm::ArrayRef<T> findSubrange(llvm::ArrayRef<U> Subrange, |
| 190 | llvm::ArrayRef<T> Range, Eq F) { |
| 191 | assert(Subrange.size() >= 1); |
| 192 | if (Range.size() < Subrange.size()) |
| 193 | return llvm::ArrayRef(Range.end(), Range.end()); |
| 194 | for (auto Begin = Range.begin(), Last = Range.end() - Subrange.size(); |
| 195 | Begin <= Last; ++Begin) { |
| 196 | auto It = Begin; |
| 197 | for (auto ItSub = Subrange.begin(); ItSub != Subrange.end(); |
| 198 | ++ItSub, ++It) { |
| 199 | if (!F(*ItSub, *It)) |
| 200 | goto continue_outer; |
| 201 | } |
| 202 | return llvm::ArrayRef(Begin, It); |
| 203 | continue_outer:; |
| 204 | } |
| 205 | return llvm::ArrayRef(Range.end(), Range.end()); |
| 206 | } |
| 207 | |
| 208 | /// Finds a subrange in \p Tokens that match the tokens specified in \p Query. |
| 209 | /// The match should be unique. \p Query is a whitespace-separated list of |
| 210 | /// tokens to search for. |
| 211 | llvm::ArrayRef<syntax::Token> |
| 212 | findTokenRange(llvm::StringRef Query, llvm::ArrayRef<syntax::Token> Tokens) { |
| 213 | llvm::SmallVector<llvm::StringRef, 8> QueryTokens; |
| 214 | Query.split(A&: QueryTokens, Separator: ' ', /*MaxSplit=*/-1, /*KeepEmpty=*/false); |
| 215 | if (QueryTokens.empty()) { |
| 216 | ADD_FAILURE() << "will not look for an empty list of tokens" ; |
| 217 | std::abort(); |
| 218 | } |
| 219 | // An equality test for search. |
| 220 | auto TextMatches = [this](llvm::StringRef Q, const syntax::Token &T) { |
| 221 | return Q == T.text(SM: *SourceMgr); |
| 222 | }; |
| 223 | // Find a match. |
| 224 | auto Found = findSubrange(Subrange: llvm::ArrayRef(QueryTokens), Range: Tokens, F: TextMatches); |
| 225 | if (Found.begin() == Tokens.end()) { |
| 226 | ADD_FAILURE() << "could not find the subrange for " << Query; |
| 227 | std::abort(); |
| 228 | } |
| 229 | // Check that the match is unique. |
| 230 | if (findSubrange(Subrange: llvm::ArrayRef(QueryTokens), |
| 231 | Range: llvm::ArrayRef(Found.end(), Tokens.end()), F: TextMatches) |
| 232 | .begin() != Tokens.end()) { |
| 233 | ADD_FAILURE() << "match is not unique for " << Query; |
| 234 | std::abort(); |
| 235 | } |
| 236 | return Found; |
| 237 | }; |
| 238 | |
| 239 | // Specialized versions of findTokenRange for expanded and spelled tokens. |
| 240 | llvm::ArrayRef<syntax::Token> findExpanded(llvm::StringRef Query) { |
| 241 | return findTokenRange(Query, Tokens: Buffer.expandedTokens()); |
| 242 | } |
| 243 | llvm::ArrayRef<syntax::Token> findSpelled(llvm::StringRef Query, |
| 244 | FileID File = FileID()) { |
| 245 | if (!File.isValid()) |
| 246 | File = SourceMgr->getMainFileID(); |
| 247 | return findTokenRange(Query, Tokens: Buffer.spelledTokens(FID: File)); |
| 248 | } |
| 249 | |
| 250 | // Data fields. |
| 251 | DiagnosticOptions DiagOpts; |
| 252 | llvm::IntrusiveRefCntPtr<DiagnosticsEngine> Diags = |
| 253 | new DiagnosticsEngine(new DiagnosticIDs, DiagOpts); |
| 254 | IntrusiveRefCntPtr<llvm::vfs::InMemoryFileSystem> FS = |
| 255 | new llvm::vfs::InMemoryFileSystem; |
| 256 | llvm::IntrusiveRefCntPtr<FileManager> FileMgr = |
| 257 | new FileManager(FileSystemOptions(), FS); |
| 258 | llvm::IntrusiveRefCntPtr<SourceManager> SourceMgr = |
| 259 | new SourceManager(*Diags, *FileMgr); |
| 260 | /// Contains last result of calling recordTokens(). |
| 261 | TokenBuffer Buffer = TokenBuffer(*SourceMgr); |
| 262 | }; |
| 263 | |
| 264 | TEST_F(TokenCollectorTest, RawMode) { |
| 265 | EXPECT_THAT(tokenize("int main() {}" ), |
| 266 | ElementsAre(Kind(tok::kw_int), |
| 267 | AllOf(HasText("main" ), Kind(tok::identifier)), |
| 268 | Kind(tok::l_paren), Kind(tok::r_paren), |
| 269 | Kind(tok::l_brace), Kind(tok::r_brace))); |
| 270 | // Comments are ignored for now. |
| 271 | EXPECT_THAT(tokenize("/* foo */int a; // more comments" ), |
| 272 | ElementsAre(Kind(tok::kw_int), |
| 273 | AllOf(HasText("a" ), Kind(tok::identifier)), |
| 274 | Kind(tok::semi))); |
| 275 | EXPECT_THAT(tokenize("int [[main() {]]}" ), |
| 276 | ElementsAre(AllOf(HasText("main" ), Kind(tok::identifier)), |
| 277 | Kind(tok::l_paren), Kind(tok::r_paren), |
| 278 | Kind(tok::l_brace))); |
| 279 | EXPECT_THAT(tokenize("int [[main() { ]]}" ), |
| 280 | ElementsAre(AllOf(HasText("main" ), Kind(tok::identifier)), |
| 281 | Kind(tok::l_paren), Kind(tok::r_paren), |
| 282 | Kind(tok::l_brace))); |
| 283 | // First token is partially parsed, last token is fully included even though |
| 284 | // only a part of it is contained in the range. |
| 285 | EXPECT_THAT(tokenize("int m[[ain() {ret]]urn 0;}" ), |
| 286 | ElementsAre(AllOf(HasText("ain" ), Kind(tok::identifier)), |
| 287 | Kind(tok::l_paren), Kind(tok::r_paren), |
| 288 | Kind(tok::l_brace), Kind(tok::kw_return))); |
| 289 | } |
| 290 | |
| 291 | TEST_F(TokenCollectorTest, Basic) { |
| 292 | std::pair</*Input*/ std::string, /*Expected*/ std::string> TestCases[] = { |
| 293 | {"int main() {}" , |
| 294 | R"(expanded tokens: |
| 295 | int main ( ) { } |
| 296 | file './input.cpp' |
| 297 | spelled tokens: |
| 298 | int main ( ) { } |
| 299 | no mappings. |
| 300 | )" }, |
| 301 | // All kinds of whitespace are ignored. |
| 302 | {"\t\n int\t\n main\t\n (\t\n )\t\n{\t\n }\t\n" , |
| 303 | R"(expanded tokens: |
| 304 | int main ( ) { } |
| 305 | file './input.cpp' |
| 306 | spelled tokens: |
| 307 | int main ( ) { } |
| 308 | no mappings. |
| 309 | )" }, |
| 310 | // Annotation tokens are ignored. |
| 311 | {R"cpp( |
| 312 | #pragma GCC visibility push (public) |
| 313 | #pragma GCC visibility pop |
| 314 | )cpp" , |
| 315 | R"(expanded tokens: |
| 316 | <empty> |
| 317 | file './input.cpp' |
| 318 | spelled tokens: |
| 319 | # pragma GCC visibility push ( public ) # pragma GCC visibility pop |
| 320 | mappings: |
| 321 | ['#'_0, '<eof>'_13) => ['<eof>'_0, '<eof>'_0) |
| 322 | )" }, |
| 323 | // Empty files should not crash. |
| 324 | {R"cpp()cpp" , R"(expanded tokens: |
| 325 | <empty> |
| 326 | file './input.cpp' |
| 327 | spelled tokens: |
| 328 | <empty> |
| 329 | no mappings. |
| 330 | )" }, |
| 331 | // Should not crash on errors inside '#define' directives. Error is that |
| 332 | // stringification (#B) does not refer to a macro parameter. |
| 333 | { |
| 334 | R"cpp( |
| 335 | a |
| 336 | #define MACRO() A #B |
| 337 | )cpp" , |
| 338 | R"(expanded tokens: |
| 339 | a |
| 340 | file './input.cpp' |
| 341 | spelled tokens: |
| 342 | a # define MACRO ( ) A # B |
| 343 | mappings: |
| 344 | ['#'_1, '<eof>'_9) => ['<eof>'_1, '<eof>'_1) |
| 345 | )" }}; |
| 346 | for (auto &Test : TestCases) |
| 347 | EXPECT_EQ(collectAndDump(Test.first), Test.second) |
| 348 | << collectAndDump(Code: Test.first); |
| 349 | } |
| 350 | |
| 351 | TEST_F(TokenCollectorTest, Locations) { |
| 352 | // Check locations of the tokens. |
| 353 | llvm::Annotations Code(R"cpp( |
| 354 | $r1[[int]] $r2[[a]] $r3[[=]] $r4[["foo bar baz"]] $r5[[;]] |
| 355 | )cpp" ); |
| 356 | recordTokens(Code: Code.code()); |
| 357 | // Check expanded tokens. |
| 358 | EXPECT_THAT( |
| 359 | Buffer.expandedTokens(), |
| 360 | ElementsAre(AllOf(Kind(tok::kw_int), RangeIs(Code.range("r1" ))), |
| 361 | AllOf(Kind(tok::identifier), RangeIs(Code.range("r2" ))), |
| 362 | AllOf(Kind(tok::equal), RangeIs(Code.range("r3" ))), |
| 363 | AllOf(Kind(tok::string_literal), RangeIs(Code.range("r4" ))), |
| 364 | AllOf(Kind(tok::semi), RangeIs(Code.range("r5" ))), |
| 365 | Kind(tok::eof))); |
| 366 | // Check spelled tokens. |
| 367 | EXPECT_THAT( |
| 368 | Buffer.spelledTokens(SourceMgr->getMainFileID()), |
| 369 | ElementsAre(AllOf(Kind(tok::kw_int), RangeIs(Code.range("r1" ))), |
| 370 | AllOf(Kind(tok::identifier), RangeIs(Code.range("r2" ))), |
| 371 | AllOf(Kind(tok::equal), RangeIs(Code.range("r3" ))), |
| 372 | AllOf(Kind(tok::string_literal), RangeIs(Code.range("r4" ))), |
| 373 | AllOf(Kind(tok::semi), RangeIs(Code.range("r5" ))))); |
| 374 | |
| 375 | auto StartLoc = SourceMgr->getLocForStartOfFile(FID: SourceMgr->getMainFileID()); |
| 376 | for (auto &R : Code.ranges()) { |
| 377 | EXPECT_THAT( |
| 378 | Buffer.spelledTokenContaining(StartLoc.getLocWithOffset(R.Begin)), |
| 379 | Pointee(RangeIs(R))); |
| 380 | } |
| 381 | } |
| 382 | |
| 383 | TEST_F(TokenCollectorTest, LocationInMiddleOfSpelledToken) { |
| 384 | llvm::Annotations Code(R"cpp( |
| 385 | int foo = [[baa^aar]]; |
| 386 | )cpp" ); |
| 387 | recordTokens(Code: Code.code()); |
| 388 | // Check spelled tokens. |
| 389 | auto StartLoc = SourceMgr->getLocForStartOfFile(FID: SourceMgr->getMainFileID()); |
| 390 | EXPECT_THAT( |
| 391 | Buffer.spelledTokenContaining(StartLoc.getLocWithOffset(Code.point())), |
| 392 | Pointee(RangeIs(Code.range()))); |
| 393 | } |
| 394 | |
| 395 | TEST_F(TokenCollectorTest, MacroDirectives) { |
| 396 | // Macro directives are not stored anywhere at the moment. |
| 397 | std::string Code = R"cpp( |
| 398 | #define FOO a |
| 399 | #include "unresolved_file.h" |
| 400 | #undef FOO |
| 401 | #ifdef X |
| 402 | #else |
| 403 | #endif |
| 404 | #ifndef Y |
| 405 | #endif |
| 406 | #if 1 |
| 407 | #elif 2 |
| 408 | #else |
| 409 | #endif |
| 410 | #pragma once |
| 411 | #pragma something lalala |
| 412 | |
| 413 | int a; |
| 414 | )cpp" ; |
| 415 | std::string Expected = |
| 416 | "expanded tokens:\n" |
| 417 | " int a ;\n" |
| 418 | "file './input.cpp'\n" |
| 419 | " spelled tokens:\n" |
| 420 | " # define FOO a # include \"unresolved_file.h\" # undef FOO " |
| 421 | "# ifdef X # else # endif # ifndef Y # endif # if 1 # elif 2 # else " |
| 422 | "# endif # pragma once # pragma something lalala int a ;\n" |
| 423 | " mappings:\n" |
| 424 | " ['#'_0, 'int'_39) => ['int'_0, 'int'_0)\n" ; |
| 425 | EXPECT_EQ(collectAndDump(Code), Expected); |
| 426 | } |
| 427 | |
| 428 | TEST_F(TokenCollectorTest, MacroReplacements) { |
| 429 | std::pair</*Input*/ std::string, /*Expected*/ std::string> TestCases[] = { |
| 430 | // A simple object-like macro. |
| 431 | {R"cpp( |
| 432 | #define INT int const |
| 433 | INT a; |
| 434 | )cpp" , |
| 435 | R"(expanded tokens: |
| 436 | int const a ; |
| 437 | file './input.cpp' |
| 438 | spelled tokens: |
| 439 | # define INT int const INT a ; |
| 440 | mappings: |
| 441 | ['#'_0, 'INT'_5) => ['int'_0, 'int'_0) |
| 442 | ['INT'_5, 'a'_6) => ['int'_0, 'a'_2) |
| 443 | )" }, |
| 444 | // A simple function-like macro. |
| 445 | {R"cpp( |
| 446 | #define INT(a) const int |
| 447 | INT(10+10) a; |
| 448 | )cpp" , |
| 449 | R"(expanded tokens: |
| 450 | const int a ; |
| 451 | file './input.cpp' |
| 452 | spelled tokens: |
| 453 | # define INT ( a ) const int INT ( 10 + 10 ) a ; |
| 454 | mappings: |
| 455 | ['#'_0, 'INT'_8) => ['const'_0, 'const'_0) |
| 456 | ['INT'_8, 'a'_14) => ['const'_0, 'a'_2) |
| 457 | )" }, |
| 458 | // Recursive macro replacements. |
| 459 | {R"cpp( |
| 460 | #define ID(X) X |
| 461 | #define INT int const |
| 462 | ID(ID(INT)) a; |
| 463 | )cpp" , |
| 464 | R"(expanded tokens: |
| 465 | int const a ; |
| 466 | file './input.cpp' |
| 467 | spelled tokens: |
| 468 | # define ID ( X ) X # define INT int const ID ( ID ( INT ) ) a ; |
| 469 | mappings: |
| 470 | ['#'_0, 'ID'_12) => ['int'_0, 'int'_0) |
| 471 | ['ID'_12, 'a'_19) => ['int'_0, 'a'_2) |
| 472 | )" }, |
| 473 | // A little more complicated recursive macro replacements. |
| 474 | {R"cpp( |
| 475 | #define ADD(X, Y) X+Y |
| 476 | #define MULT(X, Y) X*Y |
| 477 | |
| 478 | int a = ADD(MULT(1,2), MULT(3,ADD(4,5))); |
| 479 | )cpp" , |
| 480 | "expanded tokens:\n" |
| 481 | " int a = 1 * 2 + 3 * 4 + 5 ;\n" |
| 482 | "file './input.cpp'\n" |
| 483 | " spelled tokens:\n" |
| 484 | " # define ADD ( X , Y ) X + Y # define MULT ( X , Y ) X * Y int " |
| 485 | "a = ADD ( MULT ( 1 , 2 ) , MULT ( 3 , ADD ( 4 , 5 ) ) ) ;\n" |
| 486 | " mappings:\n" |
| 487 | " ['#'_0, 'int'_22) => ['int'_0, 'int'_0)\n" |
| 488 | " ['ADD'_25, ';'_46) => ['1'_3, ';'_12)\n" }, |
| 489 | // Empty macro replacement. |
| 490 | // FIXME: the #define directives should not be glued together. |
| 491 | {R"cpp( |
| 492 | #define EMPTY |
| 493 | #define EMPTY_FUNC(X) |
| 494 | EMPTY |
| 495 | EMPTY_FUNC(1+2+3) |
| 496 | )cpp" , |
| 497 | R"(expanded tokens: |
| 498 | <empty> |
| 499 | file './input.cpp' |
| 500 | spelled tokens: |
| 501 | # define EMPTY # define EMPTY_FUNC ( X ) EMPTY EMPTY_FUNC ( 1 + 2 + 3 ) |
| 502 | mappings: |
| 503 | ['#'_0, 'EMPTY'_9) => ['<eof>'_0, '<eof>'_0) |
| 504 | ['EMPTY'_9, 'EMPTY_FUNC'_10) => ['<eof>'_0, '<eof>'_0) |
| 505 | ['EMPTY_FUNC'_10, '<eof>'_18) => ['<eof>'_0, '<eof>'_0) |
| 506 | )" }, |
| 507 | // File ends with a macro replacement. |
| 508 | {R"cpp( |
| 509 | #define FOO 10+10; |
| 510 | int a = FOO |
| 511 | )cpp" , |
| 512 | R"(expanded tokens: |
| 513 | int a = 10 + 10 ; |
| 514 | file './input.cpp' |
| 515 | spelled tokens: |
| 516 | # define FOO 10 + 10 ; int a = FOO |
| 517 | mappings: |
| 518 | ['#'_0, 'int'_7) => ['int'_0, 'int'_0) |
| 519 | ['FOO'_10, '<eof>'_11) => ['10'_3, '<eof>'_7) |
| 520 | )" }, |
| 521 | {R"cpp( |
| 522 | #define NUM 42 |
| 523 | #define ID(a) a |
| 524 | #define M 1 + ID |
| 525 | M(NUM) |
| 526 | )cpp" , |
| 527 | R"(expanded tokens: |
| 528 | 1 + 42 |
| 529 | file './input.cpp' |
| 530 | spelled tokens: |
| 531 | # define NUM 42 # define ID ( a ) a # define M 1 + ID M ( NUM ) |
| 532 | mappings: |
| 533 | ['#'_0, 'M'_17) => ['1'_0, '1'_0) |
| 534 | ['M'_17, '<eof>'_21) => ['1'_0, '<eof>'_3) |
| 535 | )" }, |
| 536 | }; |
| 537 | |
| 538 | for (auto &Test : TestCases) { |
| 539 | std::string Dump = collectAndDump(Code: Test.first); |
| 540 | EXPECT_EQ(Test.second, Dump) << Dump; |
| 541 | } |
| 542 | } |
| 543 | |
| 544 | TEST_F(TokenCollectorTest, SpecialTokens) { |
| 545 | // Tokens coming from concatenations. |
| 546 | recordTokens(Code: R"cpp( |
| 547 | #define CONCAT(a, b) a ## b |
| 548 | int a = CONCAT(1, 2); |
| 549 | )cpp" ); |
| 550 | EXPECT_THAT(std::vector<syntax::Token>(Buffer.expandedTokens()), |
| 551 | Contains(HasText("12" ))); |
| 552 | // Multi-line tokens with slashes at the end. |
| 553 | recordTokens(Code: "i\\\nn\\\nt" ); |
| 554 | EXPECT_THAT(Buffer.expandedTokens(), |
| 555 | ElementsAre(AllOf(Kind(tok::kw_int), HasText("i\\\nn\\\nt" )), |
| 556 | Kind(tok::eof))); |
| 557 | // FIXME: test tokens with digraphs and UCN identifiers. |
| 558 | } |
| 559 | |
| 560 | TEST_F(TokenCollectorTest, LateBoundTokens) { |
| 561 | // The parser eventually breaks the first '>>' into two tokens ('>' and '>'), |
| 562 | // but we choose to record them as a single token (for now). |
| 563 | llvm::Annotations Code(R"cpp( |
| 564 | template <class T> |
| 565 | struct foo { int a; }; |
| 566 | int bar = foo<foo<int$br[[>>]]().a; |
| 567 | int baz = 10 $op[[>>]] 2; |
| 568 | )cpp" ); |
| 569 | recordTokens(Code: Code.code()); |
| 570 | EXPECT_THAT(std::vector<syntax::Token>(Buffer.expandedTokens()), |
| 571 | AllOf(Contains(AllOf(Kind(tok::greatergreater), |
| 572 | RangeIs(Code.range("br" )))), |
| 573 | Contains(AllOf(Kind(tok::greatergreater), |
| 574 | RangeIs(Code.range("op" )))))); |
| 575 | } |
| 576 | |
| 577 | TEST_F(TokenCollectorTest, DelayedParsing) { |
| 578 | llvm::StringLiteral Code = R"cpp( |
| 579 | struct Foo { |
| 580 | int method() { |
| 581 | // Parser will visit method bodies and initializers multiple times, but |
| 582 | // TokenBuffer should only record the first walk over the tokens; |
| 583 | return 100; |
| 584 | } |
| 585 | int a = 10; |
| 586 | |
| 587 | struct Subclass { |
| 588 | void foo() { |
| 589 | Foo().method(); |
| 590 | } |
| 591 | }; |
| 592 | }; |
| 593 | )cpp" ; |
| 594 | std::string ExpectedTokens = |
| 595 | "expanded tokens:\n" |
| 596 | " struct Foo { int method ( ) { return 100 ; } int a = 10 ; struct " |
| 597 | "Subclass { void foo ( ) { Foo ( ) . method ( ) ; } } ; } ;\n" ; |
| 598 | EXPECT_THAT(collectAndDump(Code), StartsWith(ExpectedTokens)); |
| 599 | } |
| 600 | |
| 601 | TEST_F(TokenCollectorTest, MultiFile) { |
| 602 | addFile(Path: "./foo.h" , Contents: R"cpp( |
| 603 | #define ADD(X, Y) X+Y |
| 604 | int a = 100; |
| 605 | #include "bar.h" |
| 606 | )cpp" ); |
| 607 | addFile(Path: "./bar.h" , Contents: R"cpp( |
| 608 | int b = ADD(1, 2); |
| 609 | #define MULT(X, Y) X*Y |
| 610 | )cpp" ); |
| 611 | llvm::StringLiteral Code = R"cpp( |
| 612 | #include "foo.h" |
| 613 | int c = ADD(1, MULT(2,3)); |
| 614 | )cpp" ; |
| 615 | |
| 616 | std::string Expected = R"(expanded tokens: |
| 617 | int a = 100 ; int b = 1 + 2 ; int c = 1 + 2 * 3 ; |
| 618 | file './input.cpp' |
| 619 | spelled tokens: |
| 620 | # include "foo.h" int c = ADD ( 1 , MULT ( 2 , 3 ) ) ; |
| 621 | mappings: |
| 622 | ['#'_0, 'int'_3) => ['int'_12, 'int'_12) |
| 623 | ['ADD'_6, ';'_17) => ['1'_15, ';'_20) |
| 624 | file './foo.h' |
| 625 | spelled tokens: |
| 626 | # define ADD ( X , Y ) X + Y int a = 100 ; # include "bar.h" |
| 627 | mappings: |
| 628 | ['#'_0, 'int'_11) => ['int'_0, 'int'_0) |
| 629 | ['#'_16, '<eof>'_19) => ['int'_5, 'int'_5) |
| 630 | file './bar.h' |
| 631 | spelled tokens: |
| 632 | int b = ADD ( 1 , 2 ) ; # define MULT ( X , Y ) X * Y |
| 633 | mappings: |
| 634 | ['ADD'_3, ';'_9) => ['1'_8, ';'_11) |
| 635 | ['#'_10, '<eof>'_21) => ['int'_12, 'int'_12) |
| 636 | )" ; |
| 637 | |
| 638 | EXPECT_EQ(Expected, collectAndDump(Code)) |
| 639 | << "input: " << Code << "\nresults: " << collectAndDump(Code); |
| 640 | } |
| 641 | |
| 642 | class TokenBufferTest : public TokenCollectorTest {}; |
| 643 | |
| 644 | TEST_F(TokenBufferTest, SpelledByExpanded) { |
| 645 | recordTokens(Code: R"cpp( |
| 646 | a1 a2 a3 b1 b2 |
| 647 | )cpp" ); |
| 648 | |
| 649 | // Expanded and spelled tokens are stored separately. |
| 650 | EXPECT_THAT(findExpanded("a1 a2" ), Not(SameRange(findSpelled("a1 a2" )))); |
| 651 | // Searching for subranges of expanded tokens should give the corresponding |
| 652 | // spelled ones. |
| 653 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 b1 b2" )), |
| 654 | ValueIs(SameRange(findSpelled("a1 a2 a3 b1 b2" )))); |
| 655 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3" )), |
| 656 | ValueIs(SameRange(findSpelled("a1 a2 a3" )))); |
| 657 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2" )), |
| 658 | ValueIs(SameRange(findSpelled("b1 b2" )))); |
| 659 | |
| 660 | // Test search on simple macro expansions. |
| 661 | recordTokens(Code: R"cpp( |
| 662 | #define A a1 a2 a3 |
| 663 | #define B b1 b2 |
| 664 | |
| 665 | A split B |
| 666 | )cpp" ); |
| 667 | // Ranges going across expansion boundaries. |
| 668 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split b1 b2" )), |
| 669 | ValueIs(SameRange(findSpelled("A split B" )))); |
| 670 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3" )), |
| 671 | ValueIs(SameRange(findSpelled("A split" ).drop_back()))); |
| 672 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2" )), |
| 673 | ValueIs(SameRange(findSpelled("split B" ).drop_front()))); |
| 674 | // Ranges not fully covering macro invocations should fail. |
| 675 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2" )), std::nullopt); |
| 676 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("b2" )), std::nullopt); |
| 677 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a2 a3 split b1 b2" )), |
| 678 | std::nullopt); |
| 679 | |
| 680 | // Recursive macro invocations. |
| 681 | recordTokens(Code: R"cpp( |
| 682 | #define ID(x) x |
| 683 | #define B b1 b2 |
| 684 | |
| 685 | ID(ID(ID(a1) a2 a3)) split ID(B) |
| 686 | )cpp" ); |
| 687 | |
| 688 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2" )), |
| 689 | ValueIs(SameRange(findSpelled("( B" ).drop_front()))); |
| 690 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split b1 b2" )), |
| 691 | ValueIs(SameRange(findSpelled( |
| 692 | "ID ( ID ( ID ( a1 ) a2 a3 ) ) split ID ( B )" )))); |
| 693 | // Mixed ranges with expanded and spelled tokens. |
| 694 | EXPECT_THAT( |
| 695 | Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split" )), |
| 696 | ValueIs(SameRange(findSpelled("ID ( ID ( ID ( a1 ) a2 a3 ) ) split" )))); |
| 697 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("split b1 b2" )), |
| 698 | ValueIs(SameRange(findSpelled("split ID ( B )" )))); |
| 699 | // Macro arguments |
| 700 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1" )), |
| 701 | ValueIs(SameRange(findSpelled("a1" )))); |
| 702 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a2" )), |
| 703 | ValueIs(SameRange(findSpelled("a2" )))); |
| 704 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a3" )), |
| 705 | ValueIs(SameRange(findSpelled("a3" )))); |
| 706 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2" )), |
| 707 | ValueIs(SameRange(findSpelled("ID ( a1 ) a2" )))); |
| 708 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3" )), |
| 709 | ValueIs(SameRange(findSpelled("ID ( a1 ) a2 a3" )))); |
| 710 | |
| 711 | // Empty macro expansions. |
| 712 | recordTokens(Code: R"cpp( |
| 713 | #define EMPTY |
| 714 | #define ID(X) X |
| 715 | |
| 716 | EMPTY EMPTY ID(1 2 3) EMPTY EMPTY split1 |
| 717 | EMPTY EMPTY ID(4 5 6) split2 |
| 718 | ID(7 8 9) EMPTY EMPTY |
| 719 | )cpp" ); |
| 720 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("1 2 3" )), |
| 721 | ValueIs(SameRange(findSpelled("1 2 3" )))); |
| 722 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("4 5 6" )), |
| 723 | ValueIs(SameRange(findSpelled("4 5 6" )))); |
| 724 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("7 8 9" )), |
| 725 | ValueIs(SameRange(findSpelled("7 8 9" )))); |
| 726 | |
| 727 | // Empty mappings coming from various directives. |
| 728 | recordTokens(Code: R"cpp( |
| 729 | #define ID(X) X |
| 730 | ID(1) |
| 731 | #pragma lalala |
| 732 | not_mapped |
| 733 | )cpp" ); |
| 734 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("not_mapped" )), |
| 735 | ValueIs(SameRange(findSpelled("not_mapped" )))); |
| 736 | |
| 737 | // Multiple macro arguments |
| 738 | recordTokens(Code: R"cpp( |
| 739 | #define ID(X) X |
| 740 | #define ID2(X, Y) X Y |
| 741 | |
| 742 | ID2(ID(a1), ID(a2) a3) ID2(a4, a5 a6 a7) |
| 743 | )cpp" ); |
| 744 | // Should fail, spans multiple arguments. |
| 745 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2" )), std::nullopt); |
| 746 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a2 a3" )), |
| 747 | ValueIs(SameRange(findSpelled("ID ( a2 ) a3" )))); |
| 748 | EXPECT_THAT( |
| 749 | Buffer.spelledForExpanded(findExpanded("a1 a2 a3" )), |
| 750 | ValueIs(SameRange(findSpelled("ID2 ( ID ( a1 ) , ID ( a2 ) a3 )" )))); |
| 751 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a5 a6" )), |
| 752 | ValueIs(SameRange(findSpelled("a5 a6" )))); |
| 753 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a4 a5 a6 a7" )), |
| 754 | ValueIs(SameRange(findSpelled("ID2 ( a4 , a5 a6 a7 )" )))); |
| 755 | // Should fail, spans multiple invocations. |
| 756 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 a4" )), |
| 757 | std::nullopt); |
| 758 | |
| 759 | // https://github.com/clangd/clangd/issues/1289 |
| 760 | recordTokens(Code: R"cpp( |
| 761 | #define FOO(X) foo(X) |
| 762 | #define INDIRECT FOO(y) |
| 763 | INDIRECT // expands to foo(y) |
| 764 | )cpp" ); |
| 765 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("y" )), std::nullopt); |
| 766 | |
| 767 | recordTokens(Code: R"cpp( |
| 768 | #define FOO(X) a X b |
| 769 | FOO(y) |
| 770 | )cpp" ); |
| 771 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("y" )), |
| 772 | ValueIs(SameRange(findSpelled("y" )))); |
| 773 | |
| 774 | recordTokens(Code: R"cpp( |
| 775 | #define ID(X) X |
| 776 | #define BAR ID(1) |
| 777 | BAR |
| 778 | )cpp" ); |
| 779 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("1" )), |
| 780 | ValueIs(SameRange(findSpelled(") BAR" ).drop_front()))); |
| 781 | |
| 782 | // Critical cases for mapping of Prev/Next in spelledForExpandedSlow. |
| 783 | recordTokens(Code: R"cpp( |
| 784 | #define ID(X) X |
| 785 | ID(prev good) |
| 786 | ID(prev ID(good2)) |
| 787 | #define LARGE ID(prev ID(bad)) |
| 788 | LARGE |
| 789 | )cpp" ); |
| 790 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good" )), |
| 791 | ValueIs(SameRange(findSpelled("good" )))); |
| 792 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good2" )), |
| 793 | ValueIs(SameRange(findSpelled("good2" )))); |
| 794 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("bad" )), std::nullopt); |
| 795 | |
| 796 | recordTokens(Code: R"cpp( |
| 797 | #define PREV prev |
| 798 | #define ID(X) X |
| 799 | PREV ID(good) |
| 800 | #define LARGE PREV ID(bad) |
| 801 | LARGE |
| 802 | )cpp" ); |
| 803 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good" )), |
| 804 | ValueIs(SameRange(findSpelled("good" )))); |
| 805 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("bad" )), std::nullopt); |
| 806 | |
| 807 | recordTokens(Code: R"cpp( |
| 808 | #define ID(X) X |
| 809 | #define ID2(X, Y) X Y |
| 810 | ID2(prev, good) |
| 811 | ID2(prev, ID(good2)) |
| 812 | #define LARGE ID2(prev, bad) |
| 813 | LARGE |
| 814 | )cpp" ); |
| 815 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good" )), |
| 816 | ValueIs(SameRange(findSpelled("good" )))); |
| 817 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good2" )), |
| 818 | ValueIs(SameRange(findSpelled("good2" )))); |
| 819 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("bad" )), std::nullopt); |
| 820 | |
| 821 | // Prev from macro body. |
| 822 | recordTokens(Code: R"cpp( |
| 823 | #define ID(X) X |
| 824 | #define ID2(X, Y) X prev ID(Y) |
| 825 | ID2(not_prev, good) |
| 826 | )cpp" ); |
| 827 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("good" )), |
| 828 | ValueIs(SameRange(findSpelled("good" )))); |
| 829 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("prev good" )), std::nullopt); |
| 830 | } |
| 831 | |
| 832 | TEST_F(TokenBufferTest, NoCrashForEofToken) { |
| 833 | recordTokens(Code: R"cpp( |
| 834 | int main() { |
| 835 | )cpp" ); |
| 836 | ASSERT_TRUE(!Buffer.expandedTokens().empty()); |
| 837 | ASSERT_EQ(Buffer.expandedTokens().back().kind(), tok::eof); |
| 838 | // Expanded range including `eof` is handled gracefully (`eof` is ignored). |
| 839 | EXPECT_THAT( |
| 840 | Buffer.spelledForExpanded(Buffer.expandedTokens()), |
| 841 | ValueIs(SameRange(Buffer.spelledTokens(SourceMgr->getMainFileID())))); |
| 842 | } |
| 843 | |
| 844 | TEST_F(TokenBufferTest, ExpandedTokensForRange) { |
| 845 | recordTokens(Code: R"cpp( |
| 846 | #define SIGN(X) X##_washere |
| 847 | A SIGN(B) C SIGN(D) E SIGN(F) G |
| 848 | )cpp" ); |
| 849 | |
| 850 | SourceRange R(findExpanded(Query: "C" ).front().location(), |
| 851 | findExpanded(Query: "F_washere" ).front().location()); |
| 852 | // Expanded and spelled tokens are stored separately. |
| 853 | EXPECT_THAT(Buffer.expandedTokens(R), |
| 854 | SameRange(findExpanded("C D_washere E F_washere" ))); |
| 855 | EXPECT_THAT(Buffer.expandedTokens(SourceRange()), testing::IsEmpty()); |
| 856 | } |
| 857 | |
| 858 | TEST_F(TokenBufferTest, ExpansionsOverlapping) { |
| 859 | // Object-like macro expansions. |
| 860 | recordTokens(Code: R"cpp( |
| 861 | #define FOO 3+4 |
| 862 | int a = FOO 1; |
| 863 | int b = FOO 2; |
| 864 | )cpp" ); |
| 865 | |
| 866 | llvm::ArrayRef<syntax::Token> Foo1 = findSpelled(Query: "FOO 1" ); |
| 867 | EXPECT_THAT( |
| 868 | Buffer.expansionStartingAt(Foo1.data()), |
| 869 | ValueIs(IsExpansion(SameRange(Foo1.drop_back()), |
| 870 | SameRange(findExpanded("3 + 4 1" ).drop_back())))); |
| 871 | EXPECT_THAT( |
| 872 | Buffer.expansionsOverlapping(Foo1), |
| 873 | ElementsAre(IsExpansion(SameRange(Foo1.drop_back()), |
| 874 | SameRange(findExpanded("3 + 4 1" ).drop_back())))); |
| 875 | |
| 876 | llvm::ArrayRef<syntax::Token> Foo2 = findSpelled(Query: "FOO 2" ); |
| 877 | EXPECT_THAT( |
| 878 | Buffer.expansionStartingAt(Foo2.data()), |
| 879 | ValueIs(IsExpansion(SameRange(Foo2.drop_back()), |
| 880 | SameRange(findExpanded("3 + 4 2" ).drop_back())))); |
| 881 | EXPECT_THAT( |
| 882 | Buffer.expansionsOverlapping(llvm::ArrayRef(Foo1.begin(), Foo2.end())), |
| 883 | ElementsAre(IsExpansion(SameRange(Foo1.drop_back()), _), |
| 884 | IsExpansion(SameRange(Foo2.drop_back()), _))); |
| 885 | |
| 886 | // Function-like macro expansions. |
| 887 | recordTokens(Code: R"cpp( |
| 888 | #define ID(X) X |
| 889 | int a = ID(1+2+3); |
| 890 | int b = ID(ID(2+3+4)); |
| 891 | )cpp" ); |
| 892 | |
| 893 | llvm::ArrayRef<syntax::Token> ID1 = findSpelled(Query: "ID ( 1 + 2 + 3 )" ); |
| 894 | EXPECT_THAT(Buffer.expansionStartingAt(&ID1.front()), |
| 895 | ValueIs(IsExpansion(SameRange(ID1), |
| 896 | SameRange(findExpanded("1 + 2 + 3" ))))); |
| 897 | // Only the first spelled token should be found. |
| 898 | for (const auto &T : ID1.drop_front()) |
| 899 | EXPECT_EQ(Buffer.expansionStartingAt(&T), std::nullopt); |
| 900 | |
| 901 | llvm::ArrayRef<syntax::Token> ID2 = findSpelled(Query: "ID ( ID ( 2 + 3 + 4 ) )" ); |
| 902 | EXPECT_THAT(Buffer.expansionStartingAt(&ID2.front()), |
| 903 | ValueIs(IsExpansion(SameRange(ID2), |
| 904 | SameRange(findExpanded("2 + 3 + 4" ))))); |
| 905 | // Only the first spelled token should be found. |
| 906 | for (const auto &T : ID2.drop_front()) |
| 907 | EXPECT_EQ(Buffer.expansionStartingAt(&T), std::nullopt); |
| 908 | |
| 909 | EXPECT_THAT(Buffer.expansionsOverlapping(llvm::ArrayRef( |
| 910 | findSpelled("1 + 2" ).data(), findSpelled("4" ).data())), |
| 911 | ElementsAre(IsExpansion(SameRange(ID1), _), |
| 912 | IsExpansion(SameRange(ID2), _))); |
| 913 | |
| 914 | // PP directives. |
| 915 | recordTokens(Code: R"cpp( |
| 916 | #define FOO 1 |
| 917 | int a = FOO; |
| 918 | #pragma once |
| 919 | int b = 1; |
| 920 | )cpp" ); |
| 921 | |
| 922 | llvm::ArrayRef<syntax::Token> DefineFoo = findSpelled(Query: "# define FOO 1" ); |
| 923 | EXPECT_THAT( |
| 924 | Buffer.expansionStartingAt(&DefineFoo.front()), |
| 925 | ValueIs(IsExpansion(SameRange(DefineFoo), |
| 926 | SameRange(findExpanded("int a" ).take_front(0))))); |
| 927 | // Only the first spelled token should be found. |
| 928 | for (const auto &T : DefineFoo.drop_front()) |
| 929 | EXPECT_EQ(Buffer.expansionStartingAt(&T), std::nullopt); |
| 930 | |
| 931 | llvm::ArrayRef<syntax::Token> PragmaOnce = findSpelled(Query: "# pragma once" ); |
| 932 | EXPECT_THAT( |
| 933 | Buffer.expansionStartingAt(&PragmaOnce.front()), |
| 934 | ValueIs(IsExpansion(SameRange(PragmaOnce), |
| 935 | SameRange(findExpanded("int b" ).take_front(0))))); |
| 936 | // Only the first spelled token should be found. |
| 937 | for (const auto &T : PragmaOnce.drop_front()) |
| 938 | EXPECT_EQ(Buffer.expansionStartingAt(&T), std::nullopt); |
| 939 | |
| 940 | EXPECT_THAT( |
| 941 | Buffer.expansionsOverlapping(findSpelled("FOO ; # pragma" )), |
| 942 | ElementsAre(IsExpansion(SameRange(findSpelled("FOO ;" ).drop_back()), _), |
| 943 | IsExpansion(SameRange(PragmaOnce), _))); |
| 944 | } |
| 945 | |
| 946 | TEST_F(TokenBufferTest, TokensToFileRange) { |
| 947 | addFile(Path: "./foo.h" , Contents: "token_from_header" ); |
| 948 | llvm::Annotations Code(R"cpp( |
| 949 | #define FOO token_from_expansion |
| 950 | #include "./foo.h" |
| 951 | $all[[$i[[int]] a = FOO;]] |
| 952 | )cpp" ); |
| 953 | recordTokens(Code: Code.code()); |
| 954 | |
| 955 | auto &SM = *SourceMgr; |
| 956 | |
| 957 | // Two simple examples. |
| 958 | auto Int = findExpanded(Query: "int" ).front(); |
| 959 | auto Semi = findExpanded(Query: ";" ).front(); |
| 960 | EXPECT_EQ(Int.range(SM), FileRange(SM.getMainFileID(), Code.range("i" ).Begin, |
| 961 | Code.range("i" ).End)); |
| 962 | EXPECT_EQ(syntax::Token::range(SM, Int, Semi), |
| 963 | FileRange(SM.getMainFileID(), Code.range("all" ).Begin, |
| 964 | Code.range("all" ).End)); |
| 965 | // We don't test assertion failures because death tests are slow. |
| 966 | } |
| 967 | |
| 968 | TEST_F(TokenBufferTest, MacroExpansions) { |
| 969 | llvm::Annotations Code(R"cpp( |
| 970 | #define FOO B |
| 971 | #define FOO2 BA |
| 972 | #define CALL(X) int X |
| 973 | #define G CALL(FOO2) |
| 974 | int B; |
| 975 | $macro[[FOO]]; |
| 976 | $macro[[CALL]](A); |
| 977 | $macro[[G]]; |
| 978 | )cpp" ); |
| 979 | recordTokens(Code: Code.code()); |
| 980 | auto &SM = *SourceMgr; |
| 981 | auto Expansions = Buffer.macroExpansions(FID: SM.getMainFileID()); |
| 982 | std::vector<FileRange> ExpectedMacroRanges; |
| 983 | for (auto Range : Code.ranges(Name: "macro" )) |
| 984 | ExpectedMacroRanges.push_back( |
| 985 | x: FileRange(SM.getMainFileID(), Range.Begin, Range.End)); |
| 986 | std::vector<FileRange> ActualMacroRanges; |
| 987 | for (auto Expansion : Expansions) |
| 988 | ActualMacroRanges.push_back(x: Expansion->range(SM)); |
| 989 | EXPECT_EQ(ExpectedMacroRanges, ActualMacroRanges); |
| 990 | } |
| 991 | |
| 992 | TEST_F(TokenBufferTest, Touching) { |
| 993 | llvm::Annotations Code("^i^nt^ ^a^b^=^1;^" ); |
| 994 | recordTokens(Code: Code.code()); |
| 995 | |
| 996 | auto Touching = [&](int Index) { |
| 997 | SourceLocation Loc = SourceMgr->getComposedLoc(FID: SourceMgr->getMainFileID(), |
| 998 | Offset: Code.points()[Index]); |
| 999 | return spelledTokensTouching(Loc, Tokens: Buffer); |
| 1000 | }; |
| 1001 | auto Identifier = [&](int Index) { |
| 1002 | SourceLocation Loc = SourceMgr->getComposedLoc(FID: SourceMgr->getMainFileID(), |
| 1003 | Offset: Code.points()[Index]); |
| 1004 | const syntax::Token *Tok = spelledIdentifierTouching(Loc, Tokens: Buffer); |
| 1005 | return Tok ? Tok->text(SM: *SourceMgr) : "" ; |
| 1006 | }; |
| 1007 | |
| 1008 | EXPECT_THAT(Touching(0), SameRange(findSpelled("int" ))); |
| 1009 | EXPECT_EQ(Identifier(0), "" ); |
| 1010 | EXPECT_THAT(Touching(1), SameRange(findSpelled("int" ))); |
| 1011 | EXPECT_EQ(Identifier(1), "" ); |
| 1012 | EXPECT_THAT(Touching(2), SameRange(findSpelled("int" ))); |
| 1013 | EXPECT_EQ(Identifier(2), "" ); |
| 1014 | |
| 1015 | EXPECT_THAT(Touching(3), SameRange(findSpelled("ab" ))); |
| 1016 | EXPECT_EQ(Identifier(3), "ab" ); |
| 1017 | EXPECT_THAT(Touching(4), SameRange(findSpelled("ab" ))); |
| 1018 | EXPECT_EQ(Identifier(4), "ab" ); |
| 1019 | |
| 1020 | EXPECT_THAT(Touching(5), SameRange(findSpelled("ab =" ))); |
| 1021 | EXPECT_EQ(Identifier(5), "ab" ); |
| 1022 | |
| 1023 | EXPECT_THAT(Touching(6), SameRange(findSpelled("= 1" ))); |
| 1024 | EXPECT_EQ(Identifier(6), "" ); |
| 1025 | |
| 1026 | EXPECT_THAT(Touching(7), SameRange(findSpelled(";" ))); |
| 1027 | EXPECT_EQ(Identifier(7), "" ); |
| 1028 | |
| 1029 | ASSERT_EQ(Code.points().size(), 8u); |
| 1030 | } |
| 1031 | |
| 1032 | TEST_F(TokenBufferTest, ExpandedBySpelled) { |
| 1033 | recordTokens(Code: R"cpp( |
| 1034 | a1 a2 a3 b1 b2 |
| 1035 | )cpp" ); |
| 1036 | // Expanded and spelled tokens are stored separately. |
| 1037 | EXPECT_THAT(findExpanded("a1 a2" ), Not(SameRange(findSpelled("a1 a2" )))); |
| 1038 | // Searching for subranges of expanded tokens should give the corresponding |
| 1039 | // spelled ones. |
| 1040 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("a1 a2 a3 b1 b2" )), |
| 1041 | ElementsAre(SameRange(findExpanded("a1 a2 a3 b1 b2" )))); |
| 1042 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("a1 a2 a3" )), |
| 1043 | ElementsAre(SameRange(findExpanded("a1 a2 a3" )))); |
| 1044 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("b1 b2" )), |
| 1045 | ElementsAre(SameRange(findExpanded("b1 b2" )))); |
| 1046 | |
| 1047 | // Test search on simple macro expansions. |
| 1048 | recordTokens(Code: R"cpp( |
| 1049 | #define A a1 a2 a3 |
| 1050 | #define B b1 b2 |
| 1051 | |
| 1052 | A split B |
| 1053 | )cpp" ); |
| 1054 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("A split B" )), |
| 1055 | ElementsAre(SameRange(findExpanded("a1 a2 a3 split b1 b2" )))); |
| 1056 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("A split" ).drop_back()), |
| 1057 | ElementsAre(SameRange(findExpanded("a1 a2 a3" )))); |
| 1058 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("split B" ).drop_front()), |
| 1059 | ElementsAre(SameRange(findExpanded("b1 b2" )))); |
| 1060 | |
| 1061 | // Ranges not fully covering macro expansions should fail. |
| 1062 | recordTokens(Code: R"cpp( |
| 1063 | #define ID(x) x |
| 1064 | |
| 1065 | ID(a) |
| 1066 | )cpp" ); |
| 1067 | // Spelled don't cover entire mapping (missing ID token) -> empty result |
| 1068 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("( a )" )), IsEmpty()); |
| 1069 | // Spelled don't cover entire mapping (missing ) token) -> empty result |
| 1070 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( a" )), IsEmpty()); |
| 1071 | |
| 1072 | // Recursive macro invocations. |
| 1073 | recordTokens(Code: R"cpp( |
| 1074 | #define ID(x) x |
| 1075 | #define B b1 b2 |
| 1076 | |
| 1077 | ID(ID(ID(a1) a2 a3)) split ID(B) |
| 1078 | )cpp" ); |
| 1079 | |
| 1080 | EXPECT_THAT( |
| 1081 | Buffer.expandedForSpelled(findSpelled("ID ( ID ( ID ( a1 ) a2 a3 ) )" )), |
| 1082 | ElementsAre(SameRange(findExpanded("a1 a2 a3" )))); |
| 1083 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( B )" )), |
| 1084 | ElementsAre(SameRange(findExpanded("b1 b2" )))); |
| 1085 | EXPECT_THAT(Buffer.expandedForSpelled( |
| 1086 | findSpelled("ID ( ID ( ID ( a1 ) a2 a3 ) ) split ID ( B )" )), |
| 1087 | ElementsAre(SameRange(findExpanded("a1 a2 a3 split b1 b2" )))); |
| 1088 | // FIXME: these should succeed, but we do not support macro arguments yet. |
| 1089 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("a1" )), IsEmpty()); |
| 1090 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( a1 ) a2" )), |
| 1091 | IsEmpty()); |
| 1092 | |
| 1093 | // Empty macro expansions. |
| 1094 | recordTokens(Code: R"cpp( |
| 1095 | #define EMPTY |
| 1096 | #define ID(X) X |
| 1097 | |
| 1098 | EMPTY EMPTY ID(1 2 3) EMPTY EMPTY split1 |
| 1099 | EMPTY EMPTY ID(4 5 6) split2 |
| 1100 | ID(7 8 9) EMPTY EMPTY |
| 1101 | )cpp" ); |
| 1102 | // Covered by empty expansions on one of both of the sides. |
| 1103 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( 1 2 3 )" )), |
| 1104 | ElementsAre(SameRange(findExpanded("1 2 3" )))); |
| 1105 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( 4 5 6 )" )), |
| 1106 | ElementsAre(SameRange(findExpanded("4 5 6" )))); |
| 1107 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( 7 8 9 )" )), |
| 1108 | ElementsAre(SameRange(findExpanded("7 8 9" )))); |
| 1109 | // Including the empty macro expansions on the side. |
| 1110 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("EMPTY ID ( 1 2 3 )" )), |
| 1111 | ElementsAre(SameRange(findExpanded("1 2 3" )))); |
| 1112 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("ID ( 1 2 3 ) EMPTY" )), |
| 1113 | ElementsAre(SameRange(findExpanded("1 2 3" )))); |
| 1114 | EXPECT_THAT( |
| 1115 | Buffer.expandedForSpelled(findSpelled("EMPTY ID ( 1 2 3 ) EMPTY" )), |
| 1116 | ElementsAre(SameRange(findExpanded("1 2 3" )))); |
| 1117 | |
| 1118 | // Empty mappings coming from various directives. |
| 1119 | recordTokens(Code: R"cpp( |
| 1120 | #define ID(X) X |
| 1121 | ID(1) |
| 1122 | #pragma lalala |
| 1123 | not_mapped |
| 1124 | )cpp" ); |
| 1125 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("# define ID ( X ) X" )), |
| 1126 | IsEmpty()); |
| 1127 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("# pragma lalala" )), |
| 1128 | IsEmpty()); |
| 1129 | |
| 1130 | // Empty macro expansion. |
| 1131 | recordTokens(Code: R"cpp( |
| 1132 | #define EMPTY |
| 1133 | EMPTY int a = 100; |
| 1134 | )cpp" ); |
| 1135 | EXPECT_THAT(Buffer.expandedForSpelled(findSpelled("EMPTY int" ).drop_back()), |
| 1136 | IsEmpty()); |
| 1137 | } |
| 1138 | |
| 1139 | TEST_F(TokenCollectorTest, Pragmas) { |
| 1140 | // Tokens coming from concatenations. |
| 1141 | recordTokens(Code: R"cpp( |
| 1142 | void foo() { |
| 1143 | #pragma unroll 4 |
| 1144 | for(int i=0;i<4;++i); |
| 1145 | } |
| 1146 | )cpp" ); |
| 1147 | } |
| 1148 | } // namespace |
| 1149 | |